Face discrimination is an outstanding example of human perceptual expertise, as it requires discrimination at the individual exemplar level within a highly homogeneous category of stimuli with similar global organization. Yet the ability to recognize a face seems effortless and is normally very efficient. To achieve this level of efficiency, specially-tuned perceptual processes are triggered when physiognomic qualities signal the existence of a face in the visual field. These processes include the analysis of characteristic face features (e.g. shape of the eyes, arc of the nose, etc.) as well as the computation of precise spatial metrics between the features themselves as well as their spatial relations to the outer contour of a face. Combining results reported in the literature with those found during our previous funding period, we propose that in the absence of quick and efficient basic-level categorization of visual stimuli as faces, the additional computations that allow discrimination among individual faces are either not elicited or are elicited disproportionately by face compared to objects. The consequence of such a state is impaired or inefficient face identification and, in extreme cases, face blindness, as can be observed in some individuals who never learn to identify faces (often referred to as congenital prosopagnosia or CP). One goal of the proposed experiments is to further explore individual variations in face processing among participants with normal face identification abilities as well as face identification deficits in CP and other face deficient populations. We also propose studies of face perception in selected patients with right hemisphere damage due to stroke for whom local processing can be relatively spared while global organization is impaired. We will continue to collect EEG and behavioral to study the perceptual and neural mechanisms underlying successful face perception. The studies proposed are expected to increase our understanding of the extraordinary human ability of face identification. Importantly, the results also will be useful for the development of training programs that would help to improve this socially important skill (see an example from our own research described in the preliminary studies section of this proposal).